In this edition of The Law Firm Disrupted, we ask how close the legal industry is to reliably answering a simple question: What are my true odds of winning this suit?

Welcome. I’m Roy Strom, and tell me your thoughts about the coming wave of legal analytics at

➤➤ Would you like to receive The Law Firm Disrupted straight to your inbox as a weekly email? Click here to sign up. 

How Predictable Can Litigation Get and How Fast?

High on any hypothetical list of things that can disrupt the legal industry is the ability to accurately predict and price the outcome of litigation.

The knowledge that a particular case has, say, a mere 5 percent chance of success would likely lead to less litigation. Even the angriest clients would likely cave to such great odds.

But, of course, the practice of law doesn’t work like that. At the moment, the question is how good can litigation prediction get?

BLM, an insurance-focused firm in the U.K., announced this week a partnership with the London School of Economics that will advance its goals of becoming a leader in advanced data analytics. Three professors, including experts on decision-mapping, machine learning and an actuary, will work with the firm’s proprietary data to better understand the costs and outcomes of its cases.

Reporting on this development, The Artificial Lawyer said a “litigation prediction battle” is heating up among firms in the U.K. insurance space, with BLM’s rivals such as Clyde & Co and Kennedys making similar investments.

“The idea is to move from statistics to prediction,” said Andrew Dunkley, BLM’s head of analytics.

But how good can the predictions get? Dunkley said the goal is not to get to “95 percent accuracy” when predicting a case; rather, it is to predict better than the best human lawyer or to provide new information that helps that lawyer make a better prediction.

“It becomes a weapon in the armory,” Dunkley said. “And how important that weapon is will depend on how good it is. What I am reasonably confident predicting is it will only become more important as the technology and methods get better.”

Dunkley added: “I am reasonably confident that give it 10 years and the market will not look like it does today.”

But there is an important distinction between what is going on in the U.K. insurance market and what is happening in the U.S. British firms don’t have access to a broad public data set like the PACER system maintained by U.S. federal courts. As a result, there is no private legal analytics firm like Lex Machina operating across the pond. Data analytics becomes much more firm-specific in that market.

One place I had thought to look for the leading-edge in litigation prediction is the litigation funding industry.

Their investments are made on an “underwriting” process that they say helps them value each claim. I’ve unfortunately never gotten much insight into how extensive that process is or whether it involves data-mining, although I do know most of the large funders use Lex Machina, which is very good at telling litigators how long some claim types have lasted in federal courts. But it does not do much by way of valuing litigation or measuring risk.

One partner at an Am Law 20 firm who has worked with litigation funders tells me that they are not as advanced as I might have thought. The partner, who declined to be named talking about relationships within the industry, is helping his firm use analytics to value cases.

“When I explain to [litigation funders] what I do, three of them have had the same reaction. They say, ‘You’re in the genius business,’” the partner said. They explained: “It would take a genius right now with where the analytics are to divine what the expected value of a single litigation is.”

The partner is confident, however, that the analytics will get there. One major hurdle is getting access to all of the full documents—or at least a statistically meaningful portion of them—from PACER. Those documents could also be analyzed using natural language processing to glean insights such as what type of arguments judges are most swayed by.

And if large law firms were able to reliably price the risk in a given litigation, they could offer creative payment structures. If the firm tells a client a settlement should be around $1 million, they can ask for a bonus if a deal is struck below that number or pay a penalty if it costs more.

“We’ve got a little ways to go,” the partner said. “We’re getting close. But we really need to get access to those dockets. And if we get to the day where everyone has access and can data mine dockets, then it really gets exciting. … That will reduce cases overall.”

If you are in this arena, please tell me your thoughts on how far off reliable litigation prediction is, and what the ultimate ramifications of it would be.


Roy’s Reading Corner

One reason law might not get there: There just aren’t enough smart people! Bloomberg reported on a survey of human talent in the artificial intelligence world that found there are a mere 22,000 people with the ability to create AI systems. Of those, a mere 3,000 are currently looking for jobs, while U.S. companies are trying to hire 10,000. These people are paid well: From $300,000 to $1 million or more at places like Facebook, Google and elsewhere.

Would law firms or legal technology companies be willing to match that salary? Would working in the law be as exciting as working in Silicon Valley? These are all questions I tried to ask the smart people at Elevate AI, which put out the survey. But they weren’t even interested in returning an email from a legal publication.

About that New Law-Big Law partnership: Called it? Earlier this year, I wrote about Bruce MacEwen’s prediction that a Big Law firm would partner with a New Law service provider. That appears to have happened this week, as my U.K. colleague James Booth reported about a Hogan Lovells deal that gives it access to Elevate Service’s 1,500 staff lawyers for big legal projects. (Elevate Services and Elevate AI are separate entities.)

From Booth’s report: The firm decided to partner with Los Angeles-based Elevate rather than use its own alumni or set up its own pool of contract lawyers because Elevate has an existing pool of contract lawyers, as well as significant experience with running a flexible lawyering business.

“We have used our own alumni directly, and we looked at developing our own internal pool, but decided it was better to partner with someone who focused on this full-time, which gives us immediate access to a bigger pool and means we can focus on our core business,” said Susan Bright, U.K. and Africa head at Hogan Lovells.

Quite the hangover: Last Saturday I had an outing to celebrate a friend’s engagement. I don’t know what you would call this particular celebration in the increasingly long schedule of rituals that precedes the modern wedding, but the event included asking friends to be groomsmen. With the request came a bottle of something called “Morning Recovery.

Anyway, it’s a hangover cure that apparently Big Law needs, writes ALM Intelligence’s senior analyst Nicholas Bruch. From Bruch: “There is a clear hangover after each downturn. Revenue doesn’t return to pre-downturn levels immediately. Instead, a lull sets in. In this period, revenue doesn’t necessarily contract, but it remains at a low level for several years.”

That’s it for this week. Thanks for reading all the way to the end, and if you haven’t already, here’s the sign up page to get The Law Firm Disrupted sent straight to your inbox each week.

Email with thoughts on the content of this briefing. Thank you!